skgpuppy.GaussianProcess module

class skgpuppy.GaussianProcess.GaussianProcess(x, t, cov, theta_min=None)

Bases: object

A Gaussian Process implementation based on Girard’s work. (Girard, A. Approximate Methods for Propagation of Uncertainty with Gaussian Process Models University of Glasgow, 2004) The main work is done in the Covariance class

Parameters:
  • x – inputs shape (n,d)
  • t – noisy responses
  • cov – covariance function
  • theta_min – use this vector of hyperparameters, otherwise use maximum likelihood estimation ot estimate it
estimate(x_star)

Returns an estimate of the mean and the code variance at a given point

Parameters:x_star – input vector of shape (d)
Returns:Mean and Variance at x_star
estimate_many(x_stars)
static get_realisation(x, cov, theta)

Generates a realisation of a gaussian process with the given parameters for the covariance function.

Parameters:
  • x – shape (n,d) => points where to get the realisation
  • cov – the covariance function
  • theta – parameter vector for cov
Returns:

Realisation of the gaussian process